V899 MON: AN OUTBURSTING PROTOSTAR WITH A PECULIAR LIGHT CURVE, AND ITS TRANSITION PHASES

Our long-term optical monitoring of V899 Mon started on 2009 November 30. The observations were carried out using the 2 m Himalayan Chandra Telescope (HCT) at the Indian Astronomical Observatory, Hanle (Ladakh), belonging to the Indian Institute of Astrophysics (IIA), India, and the 2 m telescope at the IUCAA (Inter-University Centre for Astronomy and Astrophysics) Girawali Observatory (IGO), Girawali (Pune), India. For optical photometry, the central 2K × 2K CCD section of the Himalaya Faint Object Spectrograph and Camera (HFOSC) with a pixel scale of 0296 was used on the HCT, and the 2K × 2K CCD of the IUCAA Faint Object Spectrograph & Camera (IFOSC) with a similar pixel scale of 03 was used on the IGO. This gave us a field of view (FOV) of ∼10 × 10 arcmin² on both HCT and IGO. A detailed description of the instruments and telescopes is available at the IAO⁸ and IGO⁹ websites. V899 Mon's field ((α,δ)₂₀₀₀ = 06^h09^m1928, −06°41' 554) was observed in standard UBVRI Bessel filters. We carried out observations for 69 nights, out of which 41 nights were observed through HCT and the remaining ones from IGO. The photometric observation log is given in Table 1. Only a portion of the table is provided here. The complete table is available in machine-readable form in the online journal.

Table 1.  Observation Log of V899 Mon

Date	JD	FWHMa	Filter(s)/Grism(s)	Instrument(s)
2009 Nov 30	2,455,166	19	V, R, gr8	HFOSC
2009 Dec 04	2,455,170	31	V, R, I	HFOSC
2009 Dec 16	2,455,182	18	V, R, I	HFOSC
2009 Dec 17	2,455,183	...	gr7, gr8	HFOSC

Note.

^aMeasured average FWHM. This is a measure of the seeing.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

The standard photometric data reduction steps like bias subtraction, median flat-fielding, and finally aperture photometry of V899 Mon were done using our publicly released¹⁰ pipeline for these instruments. Since V899 Mon source is quite bright and is not affected by any bright nebulosity, we used a 4 × FWHM aperture for optical photometry. The background sky was estimated from a ring outside the aperture radius with a width of 45. Magnitude calibration was done by solving color transformation equations for each night using Landolt's standard star field SA 98 (Landolt 1992) and the five field stars (see Figure 1) that we identified to have a stable magnitude over the period of our observations.

Zoom In Zoom Out Reset image size

Near-infrared (NIR) photometric monitoring of the source in J, H, and K/K_S bands was carried out using the HCT NIR camera (NIRCAM), the TIFR Near Infrared Spectrometer and Imager (TIRSPEC) mounted on HCT, and the TIFR Near Infrared Imaging Camera-II (TIRCAM2) mounted on the IGO telescope. NIRCAM has a 512 × 512 mercury cadmium telluride (HgCdTe) array, with a pixel size of 18 μm, which gives an FOV of ∼3.6 × 3.6 arcmin² on HCT. Filters used for observations were J (λ_center = 1.28 μm, Δλ = 0.28 μm), H (λ_center = 1.66 μm, Δλ = 0.33 μm), and K (λ_center = 2.22 μm, Δλ = 0.38 μm). Further details of the instrument are available at http://www.iiap.res.in/iao/nir.html. TIRSPEC has a 1024 × 1024 Hawaii-1 PACE¹¹ (HgCdTe) array, with a pixel size of 18 μm, which gives an FOV of ∼5 × 5 arcmin² on HCT. Filters used for observations were J (λ_center = 1.25 μm, Δλ = 0.16 μm), H (λ_center = 1.635 μm, Δλ = 0.29 μm), and K_S (λ_center = 2.145 μm, Δλ = 0.31 μm) (Mauna Kea Observatories near-infrared filter system). Further details of TIRSPEC are available in Ninan et al. (2014) and Ojha et al. (2012). TIRCAM2 has a 512 × 512 indium antimonide (InSb) array with a pixel size of 27 μm. Filters used for observations were J (λ_center = 1.20 μm, Δλ = 0.36 μm), H (λ_center = 1.66 μm, Δλ = 0.30 μm), and K (λ_center = 2.19 μm, Δλ = 0.40 μm). More details on TIRCAM2 are available in Naik et al. (2012). The log of NIR observations is listed in Table 1. We have photometric observations for a total of 25 nights in the NIR. Our first JHK observation was carried out during the peak of the first outburst on 2010 February 20, and the remaining observations were carried out during the current ongoing second outburst.

NIR observations were carried out by the standard telescope dithering procedure. Five dither positions were observed. All the dithered object frames were median combined to generate master sky frames for NIRCAM and TIRCAM2. We combined twilight flats and all non-extended source frames observed during the same night to create accurate flats for each night. For NIRCAM and TIRCAM2, data reduction and final photometry were performed using the standard tasks in IRAF, while the TIRSPEC data were reduced using our TIRSPEC photometry pipeline (Ninan et al. 2014). For aperture photometry we used an aperture of 3 × FWHM and the background sky was estimated from an annular ring outside the aperture radius with a width of 45. Magnitude calibration was done by solving color transformation equations for each night using Hunt's standard star fields AS 13 and AS 9 (Hunt et al. 1998) and the five field stars (labeled in Figure 2) that we identified to have a stable magnitude and that are consistent with 2MASS¹² over the period of our observations.

Zoom In Zoom Out Reset image size

Our medium-resolution (R ∼ 1000) optical spectroscopic monitoring of V899 Mon also started on 2009 November 30. The spectroscopic observations were carried out using both HCT and IGO. The full 2K × 4K section of the HFOSC CCD spectrograph was used for HCT observations, and the 2K × 2K IFOSC CCD spectrograph was used for IGO observations. These observations were done in the effective wavelength range of 3700–9000 Å, using grism 7 (center wavelength 5300 Å) and grism 8 (center wavelength 7200 Å). The spectral resolution obtained for grism 7 and grism 8 with the 150 μm slit at IGO and the 167 μm slit at HCT was ∼7 Å. The log of spectroscopic observations is listed in Table 1. We have 8 spectra from the first outburst phase, 4 from the short quiescent phase, and 30 from the ongoing second outburst phase.

For wavelength calibration, we have used FeNe, FeAr, and HeCu lamps. Standard IRAF tasks like APALL and APSUM were used for spectral reduction using our publicly released PyRAF¹³ -based pipeline developed for both the HFOSC and IFOSC instruments.

We acquired a high-resolution (R ∼ 37,000) spectrum of V899 Mon during its second outburst phase on 2014 December 22 using the Southern African Large Telescope High Resolution Spectrograph (SALT-HRS) (Bramall et al. 2010). Both the red arm (5490–8810 Å) and the blue arm (3674–5490 Å) of the HRS instrument were used, to take a single exposure of 3170 s. For this medium-resolution instrument mode of HRS, it uses a 223 fiber to collect light from the target source, and another fiber of the same diameter is used to sample a nearby patch of the sky. Apart from target frames, the ThAr arc lamp spectrum was also taken through the sky fiber for wavelength calibration.

All the SALT-HRS data reduction was done by writing a reduction tool in Python, making use of scikit-image (van der Walt et al. 2014), scipy (Jones et al. 2001), numpy (van der Walt et al. 2011), and astropy (Astropy Collaboration et al. 2013). The additive factor to translate observed velocities in the spectrum to heliocentric velocities was found to be +0.92 km s⁻¹ using the rvcorrect task in IRAF.

NIR (1.02–2.35 μm) spectroscopic monitoring of V899 Mon started on 2013 September 25 using TIRSPEC mounted on HCT. Depending on the seeing conditions, slits with 148 or 197 widths were used. Spectra were taken in at least two dithered positions inside the slit. Spectra of an argon lamp for wavelength calibration and a tungsten lamp for the continuum flat were also taken immediately after each observation without moving any of the filter wheels. For telluric line correction, bright NIR spectroscopic standard stars within a few degrees and similar airmasses were observed immediately after observing the source. The typical spectral resolution obtained in our spectra is R ∼ 1200. The log of spectroscopic observations is listed in Table 1. It may be noted that all 12 NIR spectral observations are taken during the current ongoing second outburst phase.

NIR spectral data were reduced using the TIRSPEC pipeline (Ninan et al. 2014). After wavelength calibration, the spectrum of V899 Mon was divided by a standard star spectrum to remove telluric lines and detector fringes seen in H and K orders' spectra. This continuum-corrected spectrum was scaled using the flux estimated from photometry to obtain the flux-calibrated spectrum. Since we do not have Y-band photometry, we interpolated I-band and J-band λf_λ flux to Y band and used the interpolated flux to scale the Y-band spectrum flux.

Continuum interferometric observation of V899 Mon at 1280 MHz with 33.3 MHz bandwidth was carried out on 2014 October 17 using the Giant Metrewave Radio Telescope (GMRT), Pune, India. GMRT consists of 30 dish antennas (each 45 m in diameter) in hybrid "Y" configuration (Swarup et al. 1991). Out of these, 25 antennas were online during our observation. The standard flux calibrator 3C 147 was observed for 15 minutes at the beginning and end of the observations. For phase calibration, the nearby Very Large Array calibrator 0607-085 was observed for 10 minutes after every 20 minutes of integration on V899 Mon. The total integration on the V899 Mon source was 4.5 hr.

CASAPY software was used for the data reduction. After careful iterative bad data flagging and gain calibration, we imaged the large FOV of 28' × 28' by dividing the field into 128 w-projection planes using the CLEAN algorithm.

We obtained imaging observations of the V899 Mon region in SPIRE¹⁴ 250, 350, and 500 μm and PACS¹⁵ 70 and 160 μm from the Herschel data archive.¹⁶ SPIRE data were available for two epochs, one on 2010 September 4 (Proposal ID: KPGT_fmotte_1) and another on 2013 March 16 (Proposal ID: OT1_rgutermu_1). PACS data of V899 Mon were only available in the later epoch. Data from both the epochs were reduced using the standard Herschel's HIPE pipeline with the latest calibration file. For point-source aperture photometry of V899 Mon we first tried the sourceExtractorSussextractor routine in HIPE on level 2 SPIRE data and the multiplePointSourceAperturePhotometry routine on PACS data. We also tried point-spread function (PSF) photometry using the daophot tool in IRAF. Since all these algorithms use an annular ring for background estimation and the background is highly nonuniform, the results were found to be very sensitive to aperture and background estimates. We finally estimated fluxes using getsources (Men'shchikov et al. 2012), which uses a linearly interpolated background estimate. These estimates were found to be consistent with the differential flux between two epochs (differential flux could be estimated more accurately since the subtracted image has a uniform background). The HIPE pipeline assumes a spectral energy distribution (SED) color, α = −1 (where F_ν = ν^α). The actual color α of V899 Mon was obtained from our flux estimates, and the corresponding color correction factor was multiplied to the flux values for obtaining color-corrected flux estimates.

We obtained imaging observations of the V899 Mon region in W1 (3.4 μm), W2 (4.6 μm), W3 (12 μm), and W4 (22 μm) bands from the Wide-field Infrared Survey Explorer (WISE) data archive (Wright et al. 2010). Photometric magnitude estimates were also available in the WISE all-sky catalog. W1 and W2 bands were observed at two epochs in 2010 March 17 and September 24, whereas W3 and W4 were observed only on 2010 March 17. Since W1 and W2 intensity maps looked severely saturated, we have used only W3 and W4 WISE catalog magnitudes for our study.

In this paper, we have also compiled all other publicly available archival data of this region at different epochs. Data were obtained from the following surveys and instruments: CRTS DR2,¹⁷  POSS survey,¹⁸  USNO-B catalog,¹⁹  IRAS survey,²⁰  DENIS survey,²¹  2MASS,²²  AKARI,²³  IRAC,²⁴ and MIPS.²⁵

An extended fan-shaped optical reflection nebula is typically seen around FUors. In the case of V899 Mon we detect only a small, faint, fan-shaped reflection nebula toward the north (see Figure 1). U, B, V, R, I, J, H, and K_S magnitudes of V899 Mon from our long-term continuous monitoring are listed in Table 2. Typical magnitude errors are less than 0.03 mag. Only a portion of the table is provided here. The complete table is available in machine-readable form in the online journal. The R-band magnitude light curve is shown in Figure 3. Effective R-band magnitudes²⁶ from CRTS are overplotted in the same graph for displaying the complete picture of the initial rise of the first outburst. Images of the V899 Mon field observed on 1953 December 29, 1983 December 29, 1989 January 9, and 2000 February 8 are available from digitized POSS-1 and POSS-2 archives. Magnitudes at the two epochs (1953 and 1989) that used the R-band filter are also shown in Figure 3. It should be noted that even though these magnitudes available in the USNO-B1.0 catalog are corrected for the non-standard bandpass filter used in the POSS-I survey, their variation seen between 1953 and 1989 might still be due to systematics in calibration. Few other outburst sources like V1647 Ori, V582 Aur, and V2492 Cyg have also shown such a sudden brief transition to quiescent phase after reaching the peak of the first outburst (Kóspál et al. 2013; Ninan et al. 2013; Semkov et al. 2013). V582 Aur and V2492 Cyg's light curves were very unstable, but V1647 Ori's light curve was very similar to V899 Mon in terms of the stability during the phase transitions.

Zoom In Zoom Out Reset image size

One possible scenario that could explain the sudden dimming of the source in 2011 is occultation by a dust clump. By taking the dust size parameter R_V and change in extinction A_V as free parameters, we tried to fit the observed change in magnitudes of V, R, I, J, H, and K_S bands²⁷ between the first outburst and quiescent phases (i.e., 2.86, 2.90, 3.06, 2.96, 2.75, and 2.36 Δ mag, respectively, in each band). Coefficients in Table 2 of Cardelli et al. (1989) were used for calculation, and no positive value of R_V could fit the observed changes in magnitudes. Hence, we can safely conclude that the quiescent phase in 2011 was due to an actual break in the outburst and is not due to dust occultation. Even though we have not considered scattered excess blue light (like in UX Ori systems) in the above analysis, an almost similar magnitude drop (∼2.8) in optical and NIR magnitudes to the pre-outburst phase makes it unlikely to be a case of dust obscuration. A stronger argument against the dust obscuration scenario comes from the observed variation in spectral line fluxes during this transition (see Section 3.4). The relative flux changes observed in various spectral lines cannot be explained by a simple change in extinction.

Figure 4 shows the NIR J – H/H – K color–color (CC) diagram. The position of V899 Mon in it shows no significant extinction to the source, and it falls on the classical T Tauri (CTT) locus. The positions in the CC diagram during first and second outbursts are also different from each other. But since we have only one NIR observation from the rapidly varying ending phase of the first outburst, it might not be a good representation of the NIR color of the first outburst. The green squares measured during the 2013–2015 period of the second outburst show small movements along the reddening vector. Such short-period, small-amplitude fluctuations during the outburst phase are also seen in other FUor/EXor objects (Ninan et al. 2013; Semkov et al. 2013; Audard et al. 2014). Unlike the quiescence in 2011, these short variations in the second outburst could be due to small dust clump occultations or minor fluctuations in the accretion. The NIR J/J – K color–magnitude diagram (CMD) (see Figure 5) also indicates the general trend of V899 Mon getting redder (bluer) when it dims (brightens).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.1.1. Outburst—Quiescent Transition Phase

In contrast to existing FUor/EXor observations in the literature, we could carry out several observations of V899 Mon during its transition from outburst phase to quiescent phase and back. Figure 6 shows the V-band magnitude light curve of the source, color-coded with V – R color in the outer ring and V – I color at the center. These colors indicate that the object was reddest during the transition stages rather than during the outburst or quiescent phases. It could imply that the extra flux component of the outburst was cooler at the beginning and became hotter as the outburst reached its peak. Since the transition phase flux was cooler than the quiescent phase, outburst flux could not have initially originated on the surface of the hot central star; rather, it might have originated in the disk. This is the first direct observational evidence to support triggering of outbursts in the disk of FUors/EXors. The color change seen in this plot could be a combined effect of both change in extinction and change in temperature. Figure 7 shows the V-, R-, and I-band light curves color-coded with an extinction invariant quantity. The y-intercept along the reddening vector in any CC diagrams, for example, C_VRI = $(V-R)-(R-I)*E(V-R)/E(R-I)$ (McGehee et al. 2004), is a weighted difference of two colors, which is invariant to change in extinction A_V. While it is not possible to completely separate out the degenerate SED slope change due to temperature and extinction change, any change in this extinction invariant color implies real change in the intrinsic SED slope (temperature). V899 Mon shows significant change in C_VRI between the end of the first outburst phase, quiescent phase, transition stage, and ongoing second outburst phase. Significant variation is also seen during the peak of the first outburst, implying that the temperatures of the source during those epochs were different than the ongoing second outburst. However, since C_VRI is a weighted difference between two colors and could not tell unambiguously whether the intrinsic slope of the SED became red or blue, Figure 7 only indicates change in the intrinsic slope of the SED (temperature). It does not clarify whether reddening of flux during the transition stage was due to temperature change alone or was due to the combined effect of a short-term increase in extinction and change in temperature.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

V899 Mon was observed in WISE 12 and 22 μm bands on 2010 March 17, within 1 month after our NIR observations in J, H, and K bands from HCT on 2010 February 20 (first outburst peak). We utilized them to estimate the conventional infrared logarithmic slope of the spectrum α (where λF_λ = λ^α) used for age classification of YSOs. During the peak outburst phase (2010 March), we obtained a slope α = −0.27 between 2.16 and 11.56 μm and α = −0.51 between 11.56 and 22 μm. These values classify V899 Mon as a flat-spectrum or an early Class II YSO (Greene et al. 1994) during its peak outburst phase.

U – B, B – V, V – R, and R – I colors of Siess et al. (2000) isochrones can be used to constrain photometrically the mass, age, and extinction of the source. This is done by finding a region in the mass, age, and extinction parameter space in which colors predicted by the Siess et al. (2000) model are consistent with all the observed optical colors of V899 Mon. For this analysis we sampled Siess et al. (2000) isochrones in its valid age range 0.01–100 Myr and mass range 0.1–7 M_⊙ in log space. Nonuniform rectangular mass versus age grids were generated for each of the U – B, B – V, V – R, and R – I colors by interpolating (using 2D Bi-spline) colors predicted by the Siess et al. (2000) isochrones. U – B and B – V colors of V899 Mon during the second outburst and V – R and R – I quiescent phase colors were used for fixing the position of V899 Mon in this four-dimensional color space. The errors of the color estimates were taken to be 0.1 mag, which correspond to a symmetric 4D Gaussian in color space. The radial basis function was used to calculate the distance to the source for each point in the mass–age grid. A contour was plotted for regions within 1σ distance from V899 Mon's position. These contours were repeatedly estimated for various values of A_V ranging from 0.5 to 8 mag (see Figure 8). As seen in the figure, A_V < 2.2 mag and A_V > 4.0 mag will make all Siess et al. (2000) isochrone colors incompatible with V899 Mon and hence can be ruled out. Owing to the correlation between age and mass, we obtain tight constraints only for A_V from this analysis. Even though the analysis was done for all possible age ranges, since V899 Mon is a heavy disk accreting source, we expect its position inside contours to be only around the region where age is ∼1–5 Myr.

Zoom In Zoom Out Reset image size

Another independent way to estimate interstellar extinction is from photometry of field stars. The region around Monoceros R2 was studied in detail using 2MASS data by Lombardi et al. (2011) (Figure 4 in their paper), and the extinction to the V899 Mon location is A_V = 2.6 mag. This value is consistent with our previous color space analysis.

Figure 9 shows the position of V899 Mon in the V/V – I optical CMD during various epochs. Siess et al. (2000) isochrones after correcting for extinction of A_V = 2.6 mag are also overplotted. Since the major component of the outburst phase flux is nonstellar in origin, the closest representation of the true position of V899 Mon might be its position during quiescent phase. The scatter forbids the estimation of age, but we get a rough estimate of the mass of V899 Mon as 2 M_⊙. However, it should be noted that this estimate is sensitive to the estimated interstellar extinction value of A_V = 2.6 mag.

Zoom In Zoom Out Reset image size

Flux estimates available of the V899 Mon over a wide wavelength range from optical to submillimeter were measured at random epochs. Since we had to use multiwavelength data of nearby epochs for the construction of the SED, we chose the epoch near the peak of the first outburst (2010 February–March), during which maximum observations were available. For fitting YSO SED models, we used the online SED Fitter tool by Robitaille et al. (2007). Flux estimates from both IRAS and PACS show that the flux of V899 Mon starts rising in the 90–200 μm range after a gradual drop of SED in the mid-infrared region (see Figure 10). No SED models in Robitaille et al. (2007) could fit this double-peaked SED. Possible scenarios of such an SED are discussed in Section 3.3.1. To avoid the second far-infrared peak while fitting the SED of V899 Mon, we used all the flux estimates above 90 μm as upper limits. For constraining the SED better we also used 70 μm flux measured during the second outburst in 2013, but to account for the change in 70 μm flux between two outbursts, we used a 1 Jy error bar (which is consistent with the flux change we have seen in 250 μm SPIRE data from two epochs). Figure 10(a) shows the SED fit obtained using V, R, I, J, H, K_S, W1, W2, W3, and W4 fluxes from 2010 observations and PACS 70 μm flux from 2013 observations.

Zoom In Zoom Out Reset image size

For fitting the SED of V899 Mon's quiescent phase, we used flux estimates from different quiescent phase epochs. Figure 10(b) shows the SED fit using V, R, I, 2MASS J, H, K_S, and IRAS 12 μm fluxes. MIPS1 (24 μm), PACS (70, 160 μm), and SPIRE (250, 350, 500 μm) fluxes were used only as upper limits.

Treating far-infrared PACS and SPIRE fluxes of 2013 to be a separate independent clump, we also fitted the SED by taking these fluxes alone (see Figure 11 and Section 3.3.2).

Zoom In Zoom Out Reset image size

Stellar, disk, and envelope parameters obtained from all three SED fittings are tabulated together in Table 3.

Table 3.  SED Fit Results of V899 Mon

Parameter	Quiescence	First Outburst	Far-IR Alone
AV (mag)	4.5 ± 1.1	4.2 ± 0.34	...
Distance (pc)	870 ± 80	794 ± 30	891 ± 55
Age* (Myr)	4.84 ± 2.1	1.93 ± 1.8	0.0073 ± .002
Mass* (M⊙)	3.7 ± 0.3	5.1 ± 0.5	0.57 ± 0.04
${\dot{M}}_{\mathrm{envelope}}$ (M⊙)	0 ± 5.7e-09	0 ± 7.9e-08	24.7e-05 ± 2.4e-05
MassDisk (M⊙)	1.2e-05 ± 0.0179	1.4e-03 ± 2.07e-03	0.048 ± 0.008
${\dot{M}}_{\mathrm{disk}}$ (M⊙ yr−1)	1.54e-10 ± 4.9e-07	4.82e-09 ± 7.6e-08	5.48e-07 ± 2.7e-07
Ltotal (L⊙)	162 ± 79	419 ± 168	8.55 ± 1.54
Massenv (M⊙)	8.13e-08 ± 0.027	1.52e-05 ± 0.07	23.43 ± 2.0

Note. Median and standard deviation of the best-fitted 20 models in each case.

Download table as:  ASCIITypeset image

3.3.1. 70 $\mu m$ Dip in SED

The SED of V899 Mon shows a dip around 70–90 μm in both IRAS and PACS data taken at two different epochs. This shape of the SED is very unusual and not seen in any other FUors/EXors. Objects with extended circumstellar envelopes typically show a smooth flat SED. We also do not see heavy extinction in optical bands as expected from a source surrounded by a large extended envelope. One possible scenario is that this object, instead of having a steady-state infall density distribution of the envelope, has a huge spherical cavity around it to a certain radius in the envelope. Significant optical light that we detect can be explained if our line of sight is along the opening created by the outflow.

Another possible geometric scenario is an envelope/clump with most of its mass behind V899 Mon. In the SPIRE 500 μm image (see Figure 12), we see that the bright blob marked with an ellipse at the position of V899 Mon is at the edge of a cloud, possibly being pushed back by the ionizing source of Monoceros R2.

Zoom In Zoom Out Reset image size

We plan to carry out a detailed 3D radiative transfer modeling of such geometric structures to see whether this seemingly two-component SED can be explained.

3.3.2. Far-infrared Component Properties

Considering the far-infrared component part of the SED as a separate clump, we analyzed PACS-SPIRE data alone to obtain the characteristics of the clump. L_bol of the clump obtained by fitting the Robitaille et al. (2007) SED model for the far-infrared flux (>70 μm) is ∼8.6 L_⊙ (Table 3).

We have SPIRE 250, 350, and 500 μm observations at two epochs: during the peak of the first outburst on 2010 September 4, and later during the second outburst in 2013 March 6. PACS 70 and 160 μm observations were available only during the second epoch. Photometric flux values estimated for V899 Mon from PACS and SPIRE data are given in Table 4. During the peak of the first outburst, fluxes were brighter than the second outburst by 2.8 Jy, 1.5 Jy, and 0.3 Jy in 250, 350, and 500 μm, respectively. V899 Mon was also optically brighter during the peak of the first outburst than the second outburst epoch by 0.75 mag (factor of 2) in R band. This implies that the far-infrared clump is heated by V899 Mon and is not spatially far from V899 Mon. We fitted the graybody model to SPIRE data from the two epochs (Figure 13) and estimated the mass to be 20 M_⊙ and temperature of the far-infrared clump to be 10.6 and 10.0 K at each epoch. This mass estimate is consistent with the envelope mass obtained from Robitaille et al. (2007) SED fitting (see Table 3). The graybody model was fitted using the following formulation:

$$\begin{eqnarray}&&{S}_{\nu }(\nu )\;=\;\displaystyle \frac{M(0.01{(\displaystyle \frac{\nu }{1000\;\mathrm{GHz}})}^{\beta })B(T,\nu )}{{D}^{2}},\end{eqnarray} \tag{ 1 }$$

where S_ν(ν) is the observed flux density, M is the mass of the clump, B(T, ν) is Planck's blackbody function for temperature T and frequency ν, D is the distance (taken to be 905 pc), and the dust opacity factor (κ_ν) was taken to be $0.01{\left(\frac{\nu }{1000\;\mathrm{GHz}}\right)}^{\beta }{{\rm{m}}}^{2}\;{\mathrm{kg}}^{-1},$ where β was fixed to be 2 (André et al. 2010).

Zoom In Zoom Out Reset image size

Table 4.  Far-infrared Fluxes of V899 Mon

Band	2010 Sep 14	2013 Mar 06	Δ Flux
PACS 70 μm	...	2.01 Jy	...
PACS 160 μm	...	7.02 Jy	...
SPIRE 250 μm	9.02 Jy	6.86 Jy	2.86 Jy
SPIRE 350 μm	6.82 Jy	5.68 Jy	1.48 Jy
SPIRE 500 μm	4.65 Jy	4.06 Jy	0.33 Jy

Note. Δ flux between two epochs was estimated by aperture photometry on difference images using apertures 48'', 60'', and 70'' for 250, 350, and 500 μm, respectively.

Download table as:  ASCIITypeset image

To obtain a lower limit of A_V, we assumed that the entire mass M is in a uniform-density sphere of diameter 2R = 50'' (FWHM of the source in the SPIRE 500 μm image), and then the mass column density to the center of the sphere should be 3M/(4πR²). Since most of the gas is molecular, we can equate this column density to N(H₂)${\mu }_{{{\rm{H}}}_{2}}{m}_{{\rm{H}}},$ where N(H₂) is the H₂ column number density, ${\mu }_{{{\rm{H}}}_{2}}$ is the mean molecular weight (taken to be 2.8), and m_H is the mass of hydrogen. Our mass estimate of 20 M_⊙ then corresponds to N(H₂) = 4.6 × 10²¹ molecules cm⁻². Using the relation $\langle N({{\rm{H}}}_{2})/{A}_{V}\rangle \;=\;0.94\times {10}^{21}$ molecules cm⁻² mag⁻¹ (Ciardi et al. 1998), we obtain an A_V to the center of the clump of 19.7 mag.

If we use PACS 160 μm flux also in the graybody fitting, then we obtain a mass of 12 M_⊙ and a temperature of the far-infrared clumps at the two epochs of 11.9 and 12.2 K. This mass corresponds to an A_V = 10.7 mag to the center of the uniform spherical clump. Both of these estimates are substantially higher than the actual A_V estimated from optical and NIR data. This implies that the heating source V899 Mon is not embedded at the center of this clump; it might be partially located on the front section of it (or any other geometry as discussed in Section 3.3.1). Even though the mass estimate has some systematic uncertainties from the distance, dust opacity factor, and graybody model assumption, they are unlikely to cause a difference of 8–15 mag in the A_V value.

V899 Mon has a rich spectrum of emission and absorption lines in optical and NIR. Figure 14 shows the flux-calibrated optical and NIR spectra taken during the second outburst phase of the source. Our spectroscopic observations in optical cover first outburst, transition, and quiescent phases, as well as the second outburst. The fluxes and equivalent widths of the detected lines at various epochs are tabulated in Table 5. The strong Ca ii IR triplet emission lines in the spectrum confirm the V899 Mon source to be a YSO, and the detection of CO (2–0) and (3–1) bandhead absorption starting at 2.29 μm confirms this source to be an outburst family of FUors/EXors. These overtone bandhead lines are seen only in giants and FUors/EXors. In FUors/EXors they are believed to be forming in the accretion-heated inner regions of the disk, where temperature is in the range 2000 K < T < 5000 K and density n_H > 10¹⁰ cm⁻³ (Kóspál et al. 2011). One notable line not detected in the V899 Mon spectrum is Brγ at 2.16 μm. While Brγ is typically found in strong accreting T Tauri stars, it is not detected in many FUors.

Zoom In Zoom Out Reset image size

3.4.1. Extinction Estimates

3.4.2. Accretion Rate

Many of the spectral line fluxes originating near the magnetosphere have been found to correlate with the accretion luminosity. The correlation is empirically calibrated in the simple form: $\mathrm{log}({L}_{\mathrm{acc}}/{L}_{\odot })\;=\;a\;\mathrm{log}({L}_{\mathrm{line}}/{L}_{\odot })+b.$ For estimating accretion luminosity from hydrogen Balmer series, He i, O i, Ca ii IR triplet, Paβ, and Paδ lines, we used the coefficients a and b from Alcalá et al. (2014). The distance to V899 Mon was taken to be 905 pc, and the observed fluxes were corrected for an extinction of A_V = 2.6 mag. To estimate the accretion rate from the accretion luminosity, we assumed a mass and age of V899 Mon of 2.5 M_⊙ and 1 Myr, respectively (consistent with the CMD). Stellar radius was then estimated to be 4 R_⊙ based on the isochrone for these values of mass and age (Siess et al. 2000). Finally, using the formula ${\dot{M}}_{\mathrm{disk}}\;=\;({L}_{\mathrm{acc}}{R}_{*}/{GM}){(1-{R}_{*}/{R}_{i})}^{-1}$ (where disk inner radius R_i ∼ 5 × R_*; Gullbring et al. 1998), we obtained accretion rates from various lines along the spectrum. No veiling correction was done for lines that are in absorption; hence, these accretion rate estimates can be slight underestimates. Figure 15 shows all the estimates of accretion rates from various lines as a function of their wavelengths. Figure 16 shows the normalized accretion rates obtained from different lines at various epochs. A clear reduction in relative accretion, by at least a factor of 2 during quiescent phase with respect to the second outburst phase, is seen in all the accretion rates obtained from different lines.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.4.3. Mass Loss

3.4.4. Outflow Temperature and Density

3.4.5. Line Profiles and Variability

3.4.5.1. Hα λ6563

Almost all prominent lines in the spectrum showed variability during our period of observations. The line that showed the most dramatic changes in line profile is Hα, which has a strong P Cygni absorption component and is believed to be formed in a magnetospheric accretion funnel. Figure 17 shows the evolution of the Hα P Cygni profile during its first outburst phase, quiescent phase, and the current ongoing second outburst phase. It is a selected representative sample of line profile plots from each epoch. The first spectrum published by Wils et al. (2009) shows no P Cygni profile in Hα. Our initial spectrum observed 13 days later during the peak of first outburst shows a CI Tau profile typically seen in many T Tauri stars (Stahler & Palla 2005). In subsequent spectra, the absorption component grows considerably, transforming the Hα line profile into a strong P Cygni profile (Figure 17). We also see complex structures in the absorption component of P Cygni in the spectrum taken on 2010 January 17. All these evolutions indicate that the outflow wind increased dramatically toward the end of first outburst. By the onset of the quiescent phase, the absorption component in P Cygni completely disappears and the Hα shows a near-symmetric emission line. We also do not see any P Cygni profile when the source was undergoing transition from the quiescent phase to the second outburst phase. P Cygni profiles, much fainter in strength compared to the peak of the first outburst, only start appearing after the full onset of the second outburst. Even though during the second outburst the outflow P Cygni profile is more stable than during the last phase of the first outburst, the spectrum taken on 2014 September 25 shows only a very weak P Cygni in Hα. These delays and short pauses indicate a nonsteady nature in the outflows from FUors/EXors. Table 5 shows the fluxes and equivalent widths obtained by fitting a two-component Gaussian to the line profiles. Our high-resolution spectrum (R ∼ 37,000) taken using SALT-HRS on 2014 December 22 resolves the Hα line and its multicomponent profile. Figure 18(a) shows the emission component of the Hα line profile, and Figure 18(b) shows the absorption component in the spectrum taken using SALT-HRS. The redshifted part of the profile peaks at +18 km s⁻¹ and has smooth wings with an extra broad component reaching up to a velocity of +420 km s⁻¹ (Figure 18(a)). On the other hand, the blue part of the profile shows complicated multicomponent velocity structures (Figure 18(b)). The outflow absorption component extends up to −722 km s⁻¹. We could see structures in absorption at −648, −568, −460, −274, −153, −100, and −26 km s⁻¹. These are typically associated with bulk motion within the outflowing gas (Stahler & Palla 2005). To understand how the outflow components evolve, it will be interesting to study the evolution of these components in velocity time space by carrying out multiepoch high-resolution spectroscopic observations.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.4.5.2. Forbidden lines [O i] λ6300, λ6363, [Fe ii] λ7155

Another important tracer of outflow is the forbidden line [O i] at 6300.304 Å. Our medium-resolution time evolution study does not show significant variation in this line, possibly owing to the fact that they originate in low-density jets/outflows. The SALT-HRS high-resolution spectrum shows a very interesting plateau profile for the [O i] λ6300.3 line (see Figure 19 for the resolved profile). Since we have not corrected our spectrum for telluric absorption, we shall not interpret the narrow absorption dips seen on the plateau structure. The profile has a strong redshifted emission at +25 km s⁻¹ with an FWHM of 21 km s⁻¹. The emission in the blueshifted part extends up to ∼−450 km s⁻¹. Figure 20 shows the other two weaker forbidden lines [O i] λ6363 and [Fe ii] λ7155 plotted over the [O i] λ6300 profile. We see an almost similar structure in [O i] λ6363.8, with a blueshifted emission extending up to −450 km s⁻¹ and a narrow peak emission at +20 km s⁻¹. The [Fe ii] λ7155 line also shows a very similar structure with a blue part extending up to ∼−500 km s⁻¹ and a peak emission at ∼+22 km s⁻¹. Since the forbidden line emissions are optically thin, their flux is directly proportional to the column density of emitting species. The plateau profile implies almost equal column density in the blueshifted outflow with a velocity gradient in the jet. Since we are detecting only the velocity component of the jets along the line of sight, the gradient can be either due to actual change in the outflow velocity or due to the geometrical projection effect of the outflow. A cone-shaped outflow in the direction of the observer can give rise to different projected velocities from different radial regions of the cone. A time evolution study of these profiles will give insight into the structure of the outflow. The redshifted emission peak is more difficult to explain. This might be originating from the envelope to disk infall shock regions on the surface of the disk, or from the tail part of a bipolar outflow that is moving away from us and emerging out of the region occulted by the disk in our line of sight.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.4.5.3. Ca ii IR Triplet λ8498, λ8542, λ8662

Ca ii IR triplet lines (λ8498.02, λ8542.09, λ8662.14) are believed to be originating inside and very near to the accretion column, resulting in their tight correlation with accretion rate (Muzerolle et al. 1998). Our continuous spectral monitoring shows that the P Cygni profile in Ca ii IR triplet lines evolved similar to Hα. The outflow component got stronger toward the end of the first outburst, then completely disappeared during the quiescent phase, and finally reappeared after the full onset of the second outburst. As seen in typical T Tauri stars and FUors/EXors, the line ratios of Ca ii IR triplet lines in the V899 Mon spectrum are found to be 1:1.01:0.77, representing an optically thick gas dominated by collisional decay. Unlike the emission components, the ratio of P Cygni absorption components of λ8542 and λ8662 measured from the high-resolution spectrum is 0.83:0.53 (=1.57:1), which is more consistent with atomic transition strengths (1.8:1). This implies that the absorption components in outflow are optically thin. These values are surprisingly similar to the P Cygni profiles of Ca ii IR triplet lines detected in the episodic winds of V1647 Ori (Ninan et al. 2014). Following the same arguments as in Ninan et al. (2014), we could estimate the Ca ii column density in the outflow to be 3.4 × 10¹² cm⁻² and the hydrogen column density N_H ∼ 3.8 × 10²⁰ cm⁻².

Figure 21 shows the resolved profiles of Ca ii IR triple lines. All three triplet lines (λ8498.02, λ8542.09, λ8662.14) show an asymmetric triangular profile with a steeper slope on the red side and a shallower slope on the blue side. The peaks of these lines are redshifted by +18.7, +21.4, and +23.2 km s⁻¹, respectively. This increase in the redshift is also seen in the line center estimated by fitting a Gaussian profile to the lines. The σ of the fitted Gaussian for each line is 1.177, 1.229, and 1.186 Å, respectively. The ratio of line widths of λ8498 and λ8542 is 0.96, which implies that λ8498 is ∼4% thinner than λ8542, as seen in many other T Tauri stars (Hamann & Persson 1992), though there is no statistically significant evidence for the peak of the former line to be larger than the latter one. Hamann & Persson (1992) attributed these ratios to substantial opacity broadening of λ8542 or to lower dispersion velocity in the deeper part of the region from where Ca ii IR triple lines originate. The velocities of the smooth broadened emission profiles extend up to ±150 km s⁻¹. Figure 22 shows the P Cygni absorption component in Ca ii λ8542 and λ8662 lines. The absorption components are quite broad and extend up to ∼−470 km s⁻¹ in both lines. The absorption component in λ8662 shows a prominent structure that extends only up to ∼−250 km s⁻¹.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.4.5.4. O i λ7773, λ8446

The absorption component of O i triplet lines at 7773 Å (Figure 23), which is not seen in the photosphere of cool stars, is believed to be formed in T Tauri stars owing to warm gas in the envelope or hot photosphere above the disk, and it is an indicator of the turbulence (Hamann & Persson 1992). The equivalent widths of this line in V899 Mon spectra show a very strong increase and then sudden decrease in strength just before the source went into the quiescent phase (see Figure 24). Owing to turbulence or disk rotational broadening of the O i triplet lines at 7771.94, 7774.17, and 7775.39 Å, our high-resolution spectrum (see Figure 23) shows a blended Gaussian absorption profile with σ = 2.43 Å (FWHM = 220 km s⁻¹). The velocities originating from turbulence or disk rotation extend up to ∼±200 km s⁻¹.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Another O i triplet at λ8446 (containing λ8446.25, λ8446.36, λ8446.76 triplet lines) is also seen in absorption in our SALT-HRS spectrum. These lines are also blended and have a combined FWHM = 174 km s⁻¹. The central position of the absorption line does not show any significant redshift and if present can be constrained to be <5 km s⁻¹. This implies that the ∼+20 km s⁻¹ (heliocentric velocity) redshift component seen in emission line profiles like the Ca ii IR triplet and forbidden lines are not due to any uncorrected peculiar velocity of V899 Mon with respect to our Sun.

3.4.5.5. Fe i λ8514, λ8387, λ8689, λ8675

Many Fe i emission lines were detected in our medium-resolution optical spectrum. All these lines were also resolved in SALT-HRS spectrum. Fe i λ8514 is a blend of two lines at 8514.07 and 8515.11 Å; their blended profile has an FWHM of 76 km s⁻¹. Fe i λ8387.777 has a +26.5 km s⁻¹ redshifted profile with an FWHM of 62.8 km s⁻¹, Fe i λ8688.625 has a +26.74 km s⁻¹ redshifted profile with an FWHM of 63.3 km s⁻¹, whereas a weaker Fe i λ8674.746 line has a +40 km s⁻¹ redshifted profile with an FWHM of 59.2 km s⁻¹. The redshifts in these lines are consistent with the redshifts we have measured in other resolved lines (for example, in Ca ii IR triplet lines and forbidden lines).

V899 Mon was not detected in the 1280 MHz GMRT radio continuum map observed during its second outburst. The local background noise σ in the map after cleaning was ∼0.1 mJy. We could detect eight other point sources in the 28' × 28' FOV map centered at V899 Mon's position. For this study, we take a 5σ ∼ 0.5 mJy to be a strict upper limit on V899 Mon's 1280 MHz flux.

The ionizing flux from the magnetospheric accretion can create an H ii region around the central source. Even though hydrogen Balmer lines detected in our optical spectra originate in the ionized region, since they are not optically thin, we cannot use them to constrain the extent of the H ii region.

If we consider the H ii region to be formed by a smooth, isothermal and isotropic stellar outflow, from our mass outflow rate and velocity estimates, we can obtain the expected flux using the following formula (Moran 1983):

$$\begin{eqnarray*}S(\nu ) & = & 1.3{\left(\displaystyle \frac{{T}_{e}}{{10}^{4}\;{\rm{K}}}\right)}^{0.1}{\left(\displaystyle \frac{\dot{M}}{{10}^{-6}{M}_{\odot }\;{{\rm{yr}}}^{-1}}\right)}^{4/3}\\ & & \times {\left(\displaystyle \frac{v}{{10}^{2}\;{\rm{km}}\;{{\rm{s}}}^{-1}}\right)}^{-4/3}{\left(\displaystyle \frac{D}{\mathrm{kpc}}\right)}^{-2}{\left(\displaystyle \frac{\nu }{{\rm{GHz}}}\right)}^{0.6}\;{\rm{mJy}},\end{eqnarray*}$$

where S(ν) is the expected flux density, T_e is the electron temperature of the outflowing ionized wind, $\dot{M}$ is the mass outflow rate, v is the terminal velocity of outflow, D is the distance to V899 Mon, and ν is the frequency of observation. From our spectroscopic estimates of these quantities, we can substitute $\dot{M}$ = 0.26 × 10⁻⁶ M_⊙ yr⁻¹, outflow velocity v = 1.5 × 10² km s⁻¹ , T_e = 0.9 × 10⁴ K, D = 0.905 kpc, and ν = 1.28 GHz. We obtain S(1.28 GHz) = 0.17 mJy, which is consistent with our observed upper limit of 0.5 mJy. Since the outflows we detected in the optical spectrum are unlikely to be an isotropic outflow, our predicted estimate of 0.17 mJy is also an upper limit.

Instead of considering a radial density profile due to outflow, since we have an estimate of density from forbidden optical lines, we can calculate an upper limit on the maximum size of a homogeneous, isotropic, spherical H ii region around V899 Mon. For an H ii region of radius R_s, using a spherical volume emission measure, we have the expression (Moran 1983)

$$\begin{eqnarray*}S(\nu ) & = & 3.444\times {10}^{-83}{\left(\displaystyle \frac{{10}^{4}\;{\rm{K}}}{{T}_{e}}\right)}^{0.35}\\ & & \times {\left(\displaystyle \frac{1\;{\rm{GHz}}}{\nu }\right)}^{0.1}{\left(\displaystyle \frac{1\;{\rm{kpc}}}{D}\right)}^{2}\displaystyle \frac{4\pi {R}_{s}^{3}}{3}{n}_{e}^{2}\end{eqnarray*}$$

where the radius of the H ii region R_s is in cm, and electron density n_e is per cm³. Taking 0.5 mJy as an upper limit on S(ν) and n_e ∼ 1 × 10⁶ cm⁻³, we obtain R_s < 20 AU. Since the flux of the ionizing Lyman continuum inside an H ii region is equal to the volume emission measure times the recombination rate, we have obtained the following relation for N_Lyc:

$$\begin{eqnarray*}{N}_{\mathrm{Lyc}} & = & 7.5487\times {10}^{43}\times \left(\displaystyle \frac{S(\nu )}{{\rm{mJy}}}\right){\left(\displaystyle \frac{{T}_{e}}{{10}^{4}\;{\rm{K}}}\right)}^{-0.45}\\ & & \times {\left(\displaystyle \frac{\nu }{1\;{\rm{GHz}}}\right)}^{0.1}{\left(\displaystyle \frac{D}{1\;{\rm{kpc}}}\right)}^{2}.\end{eqnarray*}$$

Using our upper limit of S(ν), we obtain an upper limit on the Lyman continuum flux from V899 Mon of N_Lyc < 5 × 10⁴³ photons s⁻¹.

According to standard instability models of FUors/EXors, a critical disk surface density is required to sustain the instability. Disk transition occurs from outburst phase to quiescent phase when its surface density drops below that critical value (Bell & Lin 1994; Zhu et al. 2009). Since the effective viscosity of the disk is lower during the quiescent phase than the outburst phase, the surface density increases more slowly during the quiescence than the rate at which it drained during the last outburst phase. Particularly in the case of V899 Mon when the first outburst stopped, if the disk density had drained below the critical density, then it has to spend more time in the subsequent quiescent phase to replenish the disk before undergoing a second outburst. Such a scenario is also seen in the light curve of another famous and very similar FUor candidate, V1647 Ori (Ninan et al. 2013). V582 Aur (FUor source) also showed multiple very brief quiescences during the ongoing outburst (Semkov et al. 2013). In order to explain these short-duration breaks in outbursts, we need to look for mechanisms during an outburst that can pause the accretion for a short duration before the disk gets critically drained. In such scenarios, the outburst can re-initiate as soon as the mechanism that was pausing the accretion disappears.

The average rate of change in V899 Mon's magnitude during the onset of first outburst was 0.038 mag month⁻¹; on the other hand, during the onset of second outburst it was larger than 0.15 mag month⁻¹. This difference implies that the timescales of the mechanisms that triggered the first and second outbursts are different.

Before V899 Mon transitioned to the quiescent phase, its spectroscopic observations showed an increase in the outflow activity (Section 3.4). Later the strong P Cygni profiles of outflow suddenly vanished as the object entered the quiescent phase. Since outflows are believed to be driven by magnetic fields, it is possible that some magnetic-field-related mechanism is responsible for abruptly pausing the accretion during the active outburst phase. Spectral line fluxes that are proportional to accretion, as well as the continuum flux of V899 Mon, increased suddenly during this short duration before the quiescence. In the final optical spectrum taken before the source transitioned from its first outburst phase to quiescent phase, we detect a sudden increase in equivalent width of the O i λ7773 absorption line (an indicator of turbulence). They are indicative of a highly turbulent activity around the V899 Mon source just before it transitioned to quiescent phase.

Accretion in low-mass YSOs is generally accepted to be via magnetospheric accretion funnels from disk to star. A large rate of accretion can provide a negative feedback in many ways. There are various semianalytic and numerical simulation studies in the literature on instabilities that can break and restart magnetospheric accretion (Kulkarni & Romanova 2008; Orlando et al. 2011; Blinova et al. 2015). For instance, the accretion can stop if the inner truncation radius of the disk moves outside the corotation radius (D'Angelo and Spruit 2010). In another scenario, the differential rotation between the inner accretion disk and the star can lead to inflation of the funnel, resulting in field lines opening and reconnecting, reducing accretion flow while enhancing outflow (Bouvier et al. 2003). These kinds of breaking mechanisms have the extra advantage that they can easily explain how a second outburst can restart without having to replenish the depleted disk within the short quiescence phase. Some other FUor sources like FU Ori, V1515 Cyg, and V2493 Cyg also show a short-duration dip in their light curves after they attained the initial peak of their outburst. All of them seem to indicate a complex negative feedback loop in magnetospheric accretion.

It is interesting to note that both V899 Mon and V1647 Ori (discussed in Ninan et al. 2013) spent shorter duration in the first outburst compared to their ongoing second outbursts. The reason for the second outburst being more stable than the first outburst could be because of the change in the inner disk's physical parameters by the heating or the draining of the disk during the first outburst (such as the extent of the interaction region between magnetosphere and inner disk, their coupling, inner radius of the disk, etc.). In the case of V1647 Ori, from the stability in the phase of the X-ray accretion spot, Hamaguchi et al. (2012) had shown that the location of the spot of the accretion column shock on the star did not change significantly between the first and second outbursts. Hence, if the quiescent phase of V1647 Ori was due to disruption in magnetic funnel accretion, at least the base of the magnetic funnel on the star surface was not completely disrupted.

To summarize, even though we do not have direct observational evidence to support increased magnetic activity, the outflow and turbulence signatures in the spectrum and continuum flux are consistent with instability in the magnetic accretion funnel, and the timescales are strongly inconsistent with any other instability models that depend on critical disk surface densities to switch on and off the outburst.

The empirical classification between FUors and EXors is based on the similarity in observed properties with classical FUors (FU Ori, V1515 Cyg, etc.) and EXors (EX Lup, V1118 Ori, etc.). Audard et al. (2014, p. 387) provide a detailed comparison of these objects from the literature. The ∼3 mag change in the brightness of V899 Mon during its outburst is typical of the EXors family of outbursts. On the other hand, the duration of outbursts in V899 Mon, ∼4 yr for the first outburst and greater than 3.5 yr for the ongoing second outburst, is significantly more than the typical duration EXors spend in outburst (less than 1–3 yr). Unlike V899 Mon, EXors also remain in quiescence longer than the time they spend in the outburst phase. However, V899 Mon's outburst timescales are still much less than classical FUors.

Spectroscopically, classical FUors have all optical and NIR hydrogen lines in absorption, while EXors have those lines in emission. In the case of V899 Mon, Hα and a small component of Hβ are in emission, while all the other hydrogen lines in optical and NIR are in absorption. CO bandheads starting at 2.29 μm are also in absorption. Hence, spectroscopically also V899 Mon lies between classical FUors and EXors. The L_bol estimates of V899 Mon from SED are greater than that of typical EXors, and they are less than that of classical FUors. The mass accretion rate during outburst phase is typical of EXors and is an order less than that of classical FUors.

Many of the recently discovered outbursts like V1647 Ori, OO Ser, etc., also show such intermediate properties between the classical bimodal classification of FUors and EXors. Discovery of more such intermediate-type outburst sources like V899 Mon indicates that this family of episodic accretion outbursts probably has a "continuum" distribution and not bimodal.